Testbed / MP-VPN testbed backup strategy

[lomaxfrog]

There is currently no backup system deployed in the testbed. The goal is to have as little as possible backups to do, and, instead, rebuild what broke.

In general terms, we have the following categories of items to backup:

• what is currently performed by hand: many of the hypervisors configuration, linux containers, mongodb, rabbitmq deployments.
• logs
• state such as MP-VPN state, over applications (perfsonar, host agent, orchestration)

What can be rebuilt may require a lot of down time and effort. Automation and documentation may address some of the burden, but perhaps not sufficiently.

State needs to be backup'ed because it cannot be re-computed or discovered. ENOS using MongoDB for almost everything, this may not be a difficult problem

This epic is to figure out what needs to be backup'ed up and how to do it.